The early decades of sound recording are a period of rapid change in both recording technology and performance practice. Primitive recording equipment, minimal documentation and musical pitch varying by time and location make it difficult for archivists to know the proper playback stylus size and playback speed. Using four very large data sets of both pristine and worn media, combined with a highly consistent selection methodology, this study investigates the level of accuracy attainable when selecting stylus size and determining playback speed. The analysis shows it is possible to determine, with high certainty, the correct playback stylus size per record label during this period of early sound recording. The analysis then shows that the distribution of pitch follows a Gaussian distribution of randomness; meaning, it is not possible to know with certainty the proper playback speed. Commentary discusses the cultural environment of the early recording era and the implications for archival practice today.